Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: fit stat: goodness-of-fit


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: fit stat: goodness-of-fit
Date   Fri, 14 Feb 2003 16:10:27 -0000

BLEEDBANDW@aol.com
>
> This is the first time that I have worked with the user
> generated "fit stat" program.  I wondered if anyone out
> there could offer me some insight as to why none of the R2
> measures are appearing for my "Poisson" syntax below?
>
> Thanks in advance for any help,
> Beth
>
> . poisson falls180 mdscog m210 chairfast m083 m214 m148
> mdsfalls180 m277 m278 m
> > 297 if los2groups==0, exposure (los91to180) cluster
> (facil)  nolog irr
>
> Poisson regression                                Number of
> obs   =        711
> Wald chi2(10)   =     142.61
> Log likelihood = -1674.5029                       Prob >
> chi2     =     0.0000
>
> (standard errors adjusted for clustering on facil)
>
>              |               Robust
> falls180 |        IRR   Std. Err.      z    P>|z|     [95%
> Conf. Interval]
>
> mdscog |   1.026908    .023846     1.14   0.253
> .9812187    1.074725
> m210 |   .1426514   .1045136    -2.66   0.008     .0339346
>   .5996653
> chairfast |    .718835   .1113207    -2.13   0.033
> .5306521    .9737523
> m083 |   .3330169   .1318824    -2.78   0.005     .1532395
>   .7237053
> m214 |   1.001907   .1976512     0.01   0.992     .6806227
>   1.474851
> m148 |   1.403981   .2481918     1.92   0.055     .9928604
>   1.985337
> mdsfalls180 |    2.19444   .2702449     6.38   0.000
> 1.723845    2.793502
> m277 |   .9953168    .181114    -0.03   0.979     .6967412
>   1.421842
> m278 |   1.673501   .3274209     2.63   0.008     1.140482
>   2.455635
> m297 |   1.033173   .1391611     0.24   0.809     .7934548
>   1.345314
> los91to180 | (exposure)
>
> . fitstat
>
> Measures of Fit for poisson of falls180
>
> Log-Lik Intercept Only:            .     Log-Lik Full
> Model:        -1674.503
> D(700):                     3349.006     LR(10):
>                 .
> Prob > LR:                         .
> McFadden's R2:                     .     McFadden's Adj R2:
>                 .
> Maximum Likelihood R2:             .     Cragg & Uhler's
> R2:                .
> AIC:                           4.741     AIC*n:
>          3371.006
> BIC:                       -1247.665     BIC':
>

I'm not very familiar with these models but all seems to
stem from the absence of a measure for log-likelihood intercept only
which is in turn a consequence of your specification of
-exposure()-. In short, your model means that these measures
cannot be computed.

Nick
n.j.cox@durham.ac.uk

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index